Bioselection of a Gain of Function Mutation that Enhances Adenovirus 5 Release and Improves Its Antitumoral Potency

Bioselection of a Gain of Function Mutation that Enhances

Adenovirus 5 Release and Improves Its Antitumoral Potency

Alena Gros, Jordi Martınez-Quintanilla, Cristina Puig, Sonia Guedan,David G. Mollevı, Ramon Alemany, and Manel Cascallo

Translational Research Laboratory, IDIBELL-Institut Catala d’Oncologia, L’Hospitalet de Llobregat, Barcelona, Spain

Abstract

Genetic bioselection of a mutagenized Ad5wt stock in humantumor xenografts led us to isolate AdT1, a mutant displaying alarge-plaque phenotype in vitro and an enhanced systemicantitumor activity in vivo . AdT1 phenotype correlates withan increased progeny release without affecting total viral yieldin different human tumors and cancer-associated fibroblasts.An approach combining hybrid Ad5/AdT1 recombinants andsequencing identified a truncating insertion in the endoplas-mic reticulum retention domain of the E3/19K protein (445Amutation) which relocates the protein to the plasma mem-brane and is responsible for AdT1’s enhanced release. E3/19K-445A phenotype does not correlate with the protein’s ability tointeract with MHC-I or induce apoptosis. Intracellular calciummeasurement revealed that the 445A mutation inducesextracellular Ca2+ influx, deregulating intracellular Ca2+

homeostasis and inducing membrane permeabilization, aviroporin-like function. E3/19K-445A mutants also displayenhanced antitumoral activity when injected both intratumor-ally and systemically in different models in vivo . Our resultsindicate that the inclusion of mutation 445A in tumor-selectiveadenoviruses would be a very powerful tool to enhance theirantitumor efficacy. [Cancer Res 2008;68(21):8928–37]

Introduction

Tumor-selective viruses are being engineered to treat cancer (1).Different strategies have been used to render adenovirus tumorspecific (2). Results to date have shown a good security/toxicityprofile for multiple modes of delivery. Flu-like symptomatologyhas been reported as the most common toxicity, especially afteri.v. delivery, along with transient transaminitis and, occasionally,thrombocytopenia (3). Antitumor effects have also been describedafter intratumoral or intracavital injection, and the potential forsynergy with chemotherapy is very promising (4). However,systemic antitumoral response rates range only from minimal tomoderate after systemic administration (5, 6).

One of the reasons that explain the poor response to oncolyticadenovirus after i.v. administration is the biodistribution profile ofadenovirus 5 (Ad5), a non–blood-borne virus that is quickly clearedfrom the bloodstream after i.v. administration and accumulates inthe liver within minutes due to the particular hepatic architectureand vascular system (7, 8). The presence of the host immune

system may also be determinant in the response to oncolyticadenoviruses, although its exact role is not clear, with somediscrepancies between preclinical data and clinical results (9, 10).In addition, other biological barriers within the tumor mass ( forexample, the tumor stroma matrix) also limits adenovirus oncolysisby hindering the spread of the virions (11, 12).

Improving adenovirus potency has been postulated as animportant step to overcome all these limitations (13). Oneapproach to render adenoviruses more cytotoxic is armingoncolytic adenoviruses with transgenes that are able to eliminatesurrounding cells and favor spreading of the oncolytic effect. Withthis aim, several proteins have been incorporated into theadenovirus backbone, including the adenovirus death protein(14), p53 (15), different prodrug-converting enzymes (16), andimmunomodulators (17). Although successful results have beenobtained with some of these constructs, the utility of this strategyhas been often limited by the small cloning capacity of mostoncolytic adenoviruses after the insertion of the genetic elementsrequired to confer selectivity.

Bioselection of randomly mutagenized adenovirus by repeatedpassaging in defined conditions can also be used as a strategy forthe isolation of mutants that exhibit potent and/or tumor-selectivephenotypes. This genetic selection has the potential to assign novelfunctions to viral genes. For example, new mutants of human Ad5with enhanced oncolytic activity were isolated after repeatedpassage of a randomly mutagenized pool in a human colorectalcancer cell line in vitro (18). A point mutation that induced a COOHterminal truncation in the i-leader protein, resulting in its earlyaccumulation, was shown to be essential for the mutant phenotype(18). Interestingly, the relevance of such mutations in the i-leaderprotein has been further confirmed by another study (19).

With the aim of selecting Ad5 mutants with enhanced antitumoractivity, we have extended the genetic selection approach byincluding the in vivo tumor environment as selective pressure.A mutant isolated from tumors, AdT1, displayed an increasedantitumor activity in vivo and large-plaque phenotype in vitro .Further characterization showed that AdT1 presented an improvedviral release without affecting the overall yield in both humantumor cells and cancer-associated fibroblasts. A combinedapproach including sequencing and analysis of recombinantsrevealed that the presence of a mutation in the endoplasmicreticulum retention domain of the E3/19K protein was responsiblefor the phenotype. We propose that the expression of the mutatedform of E3/19K is able to disturb intracellular Ca2+ homeostasisand ultimately create membrane lesions that allow enhanced virusrelease.

Materials and Methods

Cell lines and viruses. Human HEK293, A549, FaDu, SKMel-28, and NP-9

(20) cell lines and hamster pancreatic tumor HP-1 cells (21) were grown in

Note: Supplementary data for this article are available at Cancer Research Online(http://cancerres.aacrjournals.org/).

Requests for reprints: Manel Cascallo, IDIBELL-Institut Catala d’Oncologia, AvGran Via s/n km 2,7, L’Hospitalet de Llobregat, 08907 Barcelona, Spain. Phone: 34-93-2607463; Fax: 34-93-2607466; E-mail: [email protected].

I2008 American Association for Cancer Research.doi:10.1158/0008-5472.CAN-08-1145

Cancer Res 2008; 68: (21). November 1, 2008 8928 www.aacrjournals.org

Research Article

Research. on January 31, 2016. © 2008 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

DMEM supplemented with 5% fetal bovine serum (FBS). CAF1 cells, humancarcinoma-associated fibroblasts, were isolated and characterized in our

laboratory from an hepatic metastasis of a human colorectal carcinoma and

were maintained in DMEM-F12 supplemented with 10% FBS and BD MITO+

serum extender (Becton Dickinson). When evaluating effects of extracellularCa2+, CAF1 cells were seeded in fibronectin-coated dishes. To obtain >80%

infection, 293, A549, FaDu, DLD-1, SKMel-28, NP-9, CAF1 and HP-1 cells

were infected with 5, 20, 25, 15, 25, 30, 45, and 50 transducing units (TU) per

cell, respectively. As calcium-containing medium, a Ca2+-free medium towhich CaCl2 was added to a final concentration of 1.8 mmol/L was used.

Calcium-free medium was supplemented with 0.5 mmol/L EGTA immedi-

ately before use. Human Ad5 was obtained from American Type Culture

Collection (ATCC). AdTL (AdGFPLuc) has been previously described (22).Oligonucleotides were designed to sequence Ad5 and AdT1 entire viral

genomes.

Random mutagenesis and in vivo selection. An Ad5 stock was

randomly mutagenized with nitrous acid (23) and the 8 min–treated stock

(1,000-fold loss of virus viability; average of nine mutations per genome) was

chosen for the first round of in vivo selection. Mutations were fixed, and the

mutagenized stock was subsequently amplified and purified by passage at a

high TU per cell in A549 cells to prevent in vitro selection of mutants. For

each round of in vivo selection, 1 � 107 NP-9 cells were injected s.c. into the

flanks of male BALB/c nu/nu mice. Animal studies were performed in the

IDIBELL facility (AAALAC unit 1155) and approved by the IDIBELL’s Ethical

Committee for Animal Experimentation. Once the tumors reached

100 mm3, mice were injected with PBS, Ad5, or the mutagenized Ad5

stock at 2 � 1010 vp by tail vein. At 4 h postinjection, blood was drawn from

the tail and the titer was determined by an anti-hexon staining method.

Body weights of the animals were monitored, and tumor volume was

calculated as previously described (11). The virus mutants with superior

blood persistence were amplified from a blood sample of the mouse that

displayed the best tumor growth inhibition and were chosen for the next

round of in vivo selection. After the fourth round of bioselection, the virus

mutants were extracted from the tumor that showed the most pronounced

tumor growth inhibition. The AdT1 mutant was isolated from the resulting

tumor cell lysate by plaque assay in A549 cells.

Construction of Ad5/AdT1 recombinants and E3/19K modifiedviruses. Homologous recombination for the generation of modified viruses

was performed in yeast. The yeast replication elements and a selectable

marker were cloned into pAd5 plasmid (ref. 24; pAd5CAU). Homologous

recombination between different fragments of pAd5CAU and AdT1 genome(represented schematically in Fig. 2A) resulted in the generation of the

different Ad5/AdT1 recombinants used to map the mutation(s) responsible

for the large-plaque phenotype of AdT1. Plasmids pAdE3/19K-445A, pAdE3/19K-445ACS40, pAdE3/19K-KS, pAdE3/19K-CS40, and pAdE3/19K-del

(Supplementary Table S1) were constructed by recombination of pAd5CAU

with PCR-modified E3/19K sequences, and the incorporated modifications

were sequenced. Viruses were generated in 293, propagated in A549, andquantified using the anti-hexon staining–based method using 293 cells (24).

Briefly, this method consists in the infection of 293 cells with serial dilutions

of a virus sample. Thirty-six hours postinfection, cells were fixed with

methanol and incubated with mouse anti-hexon (2Hx-2 hybridoma, ATCC)and Alexa 488–labeled goat anti-mouse (Molecular Probes) during 1 h at

37jC. The similar late protein expression pattern showed by the viruses

(Supplementary Fig. S2B) assured this method was suitable for the titrationof the mutant viruses generated.

Virus release and production assays. 293, A549, CAF1, NP-9, FaDu,SkMel-28, DLD-1, or HP-1 cells were infected to allow 80% to 100% infection.

Three hours later, infection medium was removed and cells were washedtwice with PBS and incubated with fresh medium. At the indicated time

points a fraction of the supernatant (SN) and the cell-media suspension

(CE) were harvested. Viral titers were determined in triplicate according to

an anti-hexon staining–based method, as described above. For experimentsin which the effects of extracellular calcium were assayed, cells were

infected, and medium was removed 24 h later, the cells were washed twice

with PBS (-CaCl2/-MgCl2) and subsequently incubated with normal or Ca2+-

free medium.

In vitro cytotoxicity assay. Cytotoxity assay was performed by seeding30,000 A549 cells, 10,000 NP-9 cells, or 50,000 DLD-1 cells per well in 96-well

plates in DMEM with 5% FBS. Cells were infected with serial dilutions

starting with 90 TU/cell for A549, 250 TU/cell for NP-9, and 20 TU/cell for

DLD-1 cells. At day 5 postinfection, plates were washed with PBS andstained for total protein content (bicinchoninic acid assay, Pierce

Biotechnology) and absorbance was quantified. The TU per cell required

to produce 50% inhibition (IC50 value) was estimated from dose-response

curves by standard nonlinear regression (GraFit; Erithacus Software), usingan adapted Hill equation.

Western blot analysis. A549 cells were either mock-infected or infected

to allow >80% infection, and protein cell extracts were prepared

24 h postinfection with Iso-Hi-pH buffer [0.14 mol/L NaCl, 1 mmol/LMgCl2, 10 mmol/L Tris-HCl (pH 8.5), 0.5% Nonidet P-40]. Protein samples

(20 Ag/lane) were separated by SDS/PAGE and transferred to membranes.

Blots were probed with the following antibodies: anti-E3/19K (ref. 25;Tw1.3 that recognizes a luminal epitope of the protein kindly provided by

Dr. Yewdell, NIAID/NIH), anti-E1A (clone 13S-5, Santa Cruz Biotechnology),

anti-Ad5 virion proteins (ab6982, Abcam), anti-human poly(ADP-ribose)

polymerase (PARP; clone 551024, Becton Dickinson), or anti-adenovirusdeath protein (ADP; specific for residues 63–77 of the Ad2 ADP sequence

kindly provided by Dr. W.S.M. Wold, St. Louis University; ref. 26).

E3/19K and MHC-I membrane expression analysis by fluorescence-activated cell sorting. Cell surface expression of E3/19K and MHC class Iin virus-infected cells was analyzed in A549 cells using anti-E3/19K (Tw1.3)

and anti-human MHC-I (W6/32, ATCC), respectively. Twenty-four hours

postinfection, fluorescence profiles were obtained in triplicate by analyzing10,000 viable cells by flow cytometry (FACSCalibur, Becton Dickinson) using

488-nm laser and CellQuest Pro software (Becton Dickinson).

Propidium iodide uptake and measurement of intracellular Ca2+

concentration. A549 or CAF1 monolayers were infected to obtain 80%infection. Cells were harvested at various times postinfection. For

permeability studies, suspensions were incubated with 1 Ag/mL propidium

iodide (PI; Bender Medsystems) for 15 min at room temperature. The

number of permeable cells was determined in triplicate by flow cytometry(FACSCalibur, Becton Dickinson). To determine intracellular Ca2+ concen-

tration, cells were incubated with 10 Amol/L Fluo-3 acetoxymethyl ester

(AM) and 20 Amol/L Fura-red AM (Invitrogen) indicators, as previouslydescribed (27). The loaded A549 cells were analyzed in triplicate by flow

cytometry (FACSCalibur, Becton Dickinson).

Evaluation of antitumor efficacy in vivo . Pancreatic human NP-9

tumor xenografts were established s.c. into the flanks of BALB/c nu/nu mice

as described above. To evaluate systemic efficacy of the viruses carrying

E3/19K-445A mutation, mice with 100 mm3 tumors (n = 10 per group)

were treated with a single i.v. injection of PBS, Ad5, AdT1, or AdE3/19K-

445A (2 � 1010 vp/mouse in 150 AL PBS). The VP/TU ratio of the purified

virus stocks used for the in vivo efficacy experiments was similar as shown

in Supplementary Table S2. To evaluate intratumoral efficacy, 100 mm3

tumors (n = 10 per group) were treated with a single intratumoral injection

of PBS, Ad5, AdT1, or AdE3/19K-445A (2 � 109 vp/tumor in 25 AL PBS). To

evaluate the efficacy of E3/19K-445A mutants in an immune competent

model, HP-1 tumors were induced in the flanks of female 5-wk-old immune

competent Syrian (golden) hamsters (Mesocricetus auratus) by s.c.

inoculation of 1 � 106 cells. Once tumors reached 150 to 170 mm3,

hamsters were randomized (n = 10 per group), and a single intratumoral

injection was performed with Ad5 or AdT1 at 2.5 � 1010 vp/tumor in 25 ALPBS or with PBS. In each model, tumor sizes and animal body weights were

recorded and tumor volume and growth were calculated according to the

corresponding formulas. The two-tailed Student’s t test was used for

comparing the tumor progression in mice in the different treatment groups.Intratumoral adenovirus detection. Adenovirus immunofluorescence

in hamster tumors was performed as previously described (24). Briefly,

OCT-embedded sections from HP-1 tumors obtained at day 29 after virusadministration were treated with polyclonal anti-adenovirus (clone Ab6982,

Abcam Ltd.) and Alexa Fluor 488–labeled goat anti-rabbit (Molecular Probes)

antibodies and counterstained with 4¶,6-diamino-2-phenylindole. Stained

slides were analyzed under a fluorescent microscope (Olympus BX51).

A New Mutation that Enhances the Antitumoral Activity of Ad5

www.aacrjournals.org 8929 Cancer Res 2008; 68: (21). November 1, 2008



Results

In vivo bioselection of Ad5 mutants with enhanced oncolyticefficacy. A purified stock of wild-type human Ad5 was randomlymutated with nitrous acid as previously described (23). Themutagenized stock was amplified in human lung adenocarcinomaA549 cells and subjected to various rounds of selection in vivo inBALB/c nu/nu mice harboring s.c. human pancreatic NP-9 tumorsto select mutants with improved oncolytic potency (see MaterialsandMethods). In the fourth round of bioselection, the virus mutantswere extracted from the tumor that showed the most pronouncedgrowth inhibition and plaque-purified to isolate clones. One of theseclones, AdT1, was further characterized, and the mutationsresponsible for its phenotype are the main focus of this study.

Characterization of the phenotype of AdT1. To compare thesystemic antitumor efficacy of Ad5 and AdT1, mice with s.c. NP-9tumors were treated with a single i.v. dose of 2 � 1010 vp and thetumor growth was monitored. Mutant AdT1 showed an enhancedantitumor activity compared with Ad5 (Fig. 1A), and this correlatedwith intratumoral viral replication (not shown).

The enhanced antitumor efficacy displayed by AdT1 in vivo wasassociated with a large-plaque phenotype in vitro (Fig. 1B). Theplaques of AdT1 in A549 cells appeared earlier and were at leasttwice as large as the control Ad5 plaques. The AdT1 large-plaquephenotype could result from an increase in the viral yield or anincrease in viral release. The measurement of intracellular andextracellular virus in A549 cells showed that AdT1 was releasedmore efficiently from infected cells, whereas the total virusproduced was unaffected (Fig. 1C). Because fibroblasts releaseadenovirus less efficiently than epithelial cells and are relevanttarget cells in oncolysis of tumors with stroma, we analyzed andconfirmed the enhanced release of AdT1 in carcinoma-associatedfibroblasts (CAF1; Fig. 1C). This increase in the rate of viralrelease was also observed in 293 cells and human tumor celllines from head and neck (FaDu), melanoma (SkMel-28), pancreas(NP-9), and colorectal carcinoma (DLD-1), as well as in hamsterpancreatic tumor HP-1 cells (Fig. 1D), confirming that thisphenotype was not restricted to the cells used during thebioselection process. The enhanced release also resulted in an

Figure 1. Characterization of the phenotype of AdT1. A, NP-9 tumor xenografts were treated i.v. with PBS, Ad5, or AdT1. Points, mean percentage of tumorgrowth (n = 10); bars, SE. #, significant (P = 0.03) compared with the group treated with PBS; *, significant (P = 0.03) compared with mice injected with Ad5. B, plaquesize of Ad5 and AdT1 in A549 cells at day 7 postinfection. C, viral production and release kinetics of Ad5 and AdT1 in A549 cells (left ) and carcinoma-associatedfibroblasts (CAF1 cells; right ). Extracellular (SN) and intracellular (CE) viral content were analyzed at the time points indicated. *, significant (P = 8.9e�05 andP = 0.01 for A549 and CAF1 cells, respectively) compared with Ad5. D, viral release of Ad5 and AdT1 in a panel of different tumor cell lines. Titer of the supernatantat the time point at which AdT1 and Ad5 showed biggest differences for each cell line. E, comparative cytotoxicity of Ad5 and AdT1 in A549, NP-9, and DLD-1 tumorcell lines.

Cancer Research




enhanced cytotoxicity, as tested in A549, NP-9, and DLD-1 cells(Fig. 1E). The dominance of the phenotype of AdT1 wasconfirmed by performing a cell-plaque assay after coinfectionwith a GFP-expressing E1A-deleted adenoviral vector (Supple-mentary Fig. S1).

Identification of the mutation(s) responsible for theenhanced viral release of AdT1: E3/19K-445A mutation. Wecarried out the functional mapping of the AdT1 phenotype andcomplete sequence analysis of AdT1 genome. The Ad5/AdT1recombinants mapped the mutation responsible for the large-plaque phenotype to nucleotides 27335 to 35935 of AdT1(schematically represented in Fig. 2A). The sequence analysis ofthe genome revealed four mutations (Fig. 2A). Among these, twowere silent and one inserted three nucleotides (GGA) in position26296 of Ad5, which introduced an extra Gly at position 745 of theamino acid sequence of the L4-100K protein. The fourth mutation,found in position 29174, inserted an adenine into position 445 ofthe nucleotide sequence of the E3/19K protein. Because mutationE3/19K-445A was the only change detected in the smallestfragment of AdT1 that was able to confer large-plaque phenotype,we hypothesized it was responsible for the phenotype. Thegeneration of mutant AdE3/19K-445A proved that this mutationwas sufficient to enhance the release of Ad5. This mutationconferred both a large-plaque phenotype in A549 (Fig. 2B)and an enhanced viral release without modifying the viral yieldin CAF1 cells (Fig. 2C). With these data, we concluded that

mutation E3/19K-445A was responsible for the enhanced releaseobserved with AdT1.

The change in localization of the E3/19K protein enhancesadenovirus release. E3/19K protein is a small glycoprotein codedby the early region E3 of Ad5. The cytoplasmic tail of this protein,affected by mutation E3/19K-445A, contains a dilysine motifessential for the retention of E3/19K in the endoplasmic reticulum(ER; refs. 25, 28). Mutation E3/19K-445A generates a shift in theopen reading frame (ORF) that changes the amino acid sequencefrom position 149 and generates a stop codon at position 154. As aresult, E3/19K-445A presents a truncated cytoplasmic tail, whichresults in the loss of the ER retention signal and is expected torelocalize E3/19K to the plasma membrane.

The enhanced viral release observed with AdE3/19K-445A couldbe due to the loss of function of the E3/19K protein or to the gainof function of this protein when translocated to the cell surface.Two viruses were constructed to test this hypothesis: AdE3/19K-del, which lacks E3/19K expression (nucleotides 28846–29201 ofAd5 were deleted), and AdE3/19K-KS, a mutant different fromAdE3/19K-445A which also affects the ER retention domain of thisprotein (the dilysine motif was substituted for two serines, KK!SS;ref. 29; Supplementary Table S1). As Fig. 3A displays, cell surfaceexpression of the E3/19K was only detected in the cells infectedwith the viruses that had lost the ER retention signal: AdT1, AdE3/19K-445A, and AdE3/19K-KS. The lack of E3/19K expression inmutant AdE3/19K-del was confirmed by Western blot, as shown in

Figure 2. Identification of the mutations responsible for the AdT1 phenotype. A, schematic representation of the Ad5/AdT1 recombinants and the mutationspresent in the AdT1 genome. The fragments of the Ad5 (black ) and AdT1 (white) genomes and the enzymes used for the construction of the recombinants arerepresented, as well as their ability to form large plaques in A549 monolayers. The site and genes affected by the mutations are specified. Nucleotide changes (!) andinsertions (+) are indicated. Mutations located between two protein coding regions are indicated in parentheses. pTP, preterminal protein. The amino acid (aa )positions affected by the mutation are indicated in parentheses. B, plaque morphology of Ad5, AdT1 and AdE3/19K-445A in A549 cells 7 d after infection. C, viralrelease and production of Ad5, AdT1, and AdE3/19K-445A in carcinoma-associated fibroblasts. CAF1 cells were infected, and virus present in the supernatant (SN) andcell extract (CE) was measured 48 h postinfection.





Fig. 3B . Interestingly, a viral release assay in CAF1 cells showed thatAdE3/19K-KS was released 100 times more efficiently than Ad5(Fig. 3C). The phenocopy of AdT1 and AdE3/19K-445A by AdE3/19K-KS established a direct association between the change oflocalization of E3/19K and the enhanced rate of viral release andsuggested a new function for E3/19K at the cell membrane. Bycontrast, AdE3/19K-del showed levels of extracellular viruscomparable with Ad5, in concordance with previous studies statingthat the lack of E3/19K expression does not increase adenovirusrelease (30).

The enhanced release phenotype of mutant AdE3/19K-445Ais not mediated by interaction with MHC-I at the cell surfaceor by apoptosis. The main function of E3/19K is to bind to andretain the heavy chains of MHC class I in the ER, preventing CTLresponse against adenoviral infection (25, 28, 31). Consequently, theloss of ER retention of E3/19K-445A inhibits ER retention of MHC-I.We hypothesized that the interaction of E3/19K and MHC-I at theplasma membrane could be triggering the enhanced releaseobserved with AdE3/19K-445A. For this purpose, we constructedmutant AdE3/19K-445ACS40, which contained both the E3/19K-445A mutation and the substitution of a cysteine in position 40 ofthe amino acid sequence of E3/19K for a serine (C!S-40), which

prevents MHC-I binding (32). As Fig. 4A displays, the analysis ofcell surface expression of MHC-I showed that AdE3/19K-445Ablocked MHC-I transport to the plasma membrane less efficiently,although it did not reach basal levels of MHC-I expression, aspreviously described (33). On the other hand, mutant AdE3/19K-445ACS40 showed similar cell surface MHC-I expression as theuninfected cells. This confirms that this mutant is unable toprevent MHC-I transport to the membrane, because it is unable tobind to MHC-I (significant differences compared with AdE3/19K-445A). A viral release assay in CAF1 cells showed that, althoughAdE3/19K-445ACS40 was unable to bind to MHC-I, it was releasedearlier into the supernatant (Fig. 4B). This mutant also displayeda large-plaque phenotype in A549 (Supplementary Table S1),proving that the phenotype was not dependent on interaction withMHC-I. Additional evidence suggesting that the AdT1 mutant didnot enhance viral release by interacting with MHC-I comes fromthe enhanced release phenotype of AdT1 in a cell line negative forMHC-I expression, such as DLD-1 (Fig. 1D and SupplementaryTable S1).

Adenovirus mutants defective in viral genes that inhibitapoptosis produce a characteristic phenotype similar to thatobserved with AdT1 and AdE3/19K-445A (34).To determine thelevels of apoptosis activation with our mutants, we analyzed PARPcleavage during the viral cycle. As Fig. 4C displays, the level of PARPcleavage in AdT1 and AdE3/19K-445A–infected cells was similar tothat seen for Ad5, which shows that these mutants were unable toactivate apoptosis. Further confirmation of the apoptosis indepen-dence of the AdE319K-445A phenotype was gained from viralrelease experiments in CAF1 cells treated with the broad spectrumcaspase inhibitor Q-VD-OPh. Both the viral release and plaque-sizeof AdE3/19K-445A were not affected by the presence of Q-VD-OPh,proving that the enhanced rate of viral release was not mediated byapoptosis (Fig. 4D ; Supplementary Table S1).

The enhanced release of E3/19K-445A–expressing mutantsis not mediated by ADP overexpression. ADP is a smallglycoprotein coded by the E3 region of Ad5 required for theefficient lysis of adenovirus-infected cells. Its overexpression resultsin an enhanced release and large-plaque phenotype (14, 30, 35). Theexpression of this protein is tightly regulated by pre-mRNAprocessing and some deletion mutations upstream of ADP whichaffect splicing or polyadenylation sites increase the expression ofits early transcripts (26, 36). The proximity of E3/19K and ADP ORFled us to study a possible ADP overexpression caused by the E3/19K-445A mutation. No differences were found in the pattern of E3transcripts (Supplementary Fig. S2A). Early ADP transcripts (bandsd and e) were not detected 8 hours postinfection for neither Ad5nor AdT1 but were greatly amplified 24 hours postinfection (bandsd ¶ and e ¶), as previously described (26). Western blot detection ofADP expression in the different mutants confirmed these resultsand dismissed ADP overexpression in AdT1 and AdE3/19K-445A–infected cells (Supplementary Fig. S2B). Furthermore, no differ-ences in E1A nor adenovirus late protein expression were found forthe different virus mutants (Supplementary Fig. S2B).

Mutant AdE3/19K-445A enhances cell permeability andincreases intracellular [Ca2+]. Some viruses actively enhance cellpermeability during viral infection by the expression of viroporins,facilitating the influx of extracellular ions, which can enhance viralrelease through yet unknown mechanisms (37, 38). To study aputative viroporin-like function of E3/19K-445A, we analyzed cellpermeability and intracellular Ca2+ concentration throughout theviral cycle. AdE3/19K-445A enhanced cell permeability at earlier

Figure 3. Association of E3/19K localization to the plasma membrane and thephenotype of enhanced release. A, cell surface expression of E3/19K protein inA549 cells. Expression of E3/19K in the plasma membrane was analyzed at24 h postinfection in nonpermeabilized cells by flow cytometry with Tw1.3antibody. B, Western blot detection of E3/19K. A549 cell lysates were processedfor E3/19K protein expression 24 h postinfection with the adenovirus mutantsindicated. C, viral release of Ad5, AdT1, AdE3/19K-445A, AdE3/19K-KS, andAdE3/19K-del in carcinoma-associated fibroblasts. Virus present in thesupernatant was measured 48 h postinfection in CAF1.

Cancer Research




stages of infection (Fig. 5A), and this increase in cell permeabilitywas associated with an increase in the intracellular Ca2+

concentration (Fig. 5B). This enhanced permeability for AdE3/19K-445A–infected cells was also observed in CAF1 (Fig. 5C). Totest whether the influx of Ca2+ was the cause of the enhanced rateof viral release, we performed a viral release and production assayin CAF1 cells in the presence or absence of extracellular Ca2+.Although the lack of Ca2+ in the extracellular medium affected theviral yield of both viruses, the differences in viral release betweenAd5 and AdE3/19K-445A disappeared in the absence of Ca2+

(Fig. 5D), proving that the enhanced release of AdE3/19K-445A wastriggered, at least in part, by Ca2+ influx.

Evaluation of in vivo efficacy. Mice carrying NP-9 pancreatictumors were treated with a single i.v. dose of Ad5 or AdE3/19K-445A at 2 � 1010 vp/mouse. The enhanced antitumor activityshown by AdE3/19K-445A compared with Ad5 proves that the 445A

mutation is not only responsible for the enhanced release andcytotoxicity of AdT1 in vitro but also for the gain of oncolytic effectin vivo (Fig. 6A). In a second model, we assessed the comparativein vivo efficacy of Ad5, AdT1, and AdE3/19K-445A with a singleintratumoral injection of virus into NP-9 s.c. tumors. Even in suchconditions, in which we potentiate the amount of Ad5 in the tumor,both the AdT1 and AdE3/19K-445A–treated tumors respondedbetter with respect to Ad5 (Fig. 6B). We also performed a thirdin vivo experiment in Syrian hamsters that has been described asan immunocompetent model wherein human adenovirus is able toreplicate (Fig. 6C). Hamsters with s.c. pancreatic HP-1 tumors weretreated intratumorally at a dose of 2.5 � 1010 vp/tumor. Both Ad5and AdT1 displayed significant antitumoral activity, although nostatistical significance among these groups was observed. Analysisof HP-1 tumor sections revealed that despite the presence ofextensive areas of necrosis in all experimental groups, tumors

Figure 4. MHC-I binding and apoptosis independence of the enhanced release phenotype of AdT1. A, fluorescence-activated cell sorting analysis of MHC-I cellsurface expression in mutant and Ad5-infected human A549 cells. Cells were infected and, 24 h later, exposed to the monoclonal antibody W6/32. Cell surfacelevels of class I antigens are expressed relative to that of mock-infected cells, which were set to 100%. B, viral release assay of Ad5, AdE3/19K-445A, andAdE3/19K-445ACS40 in carcinoma-associated fibroblasts. Cells were infected, and the virus present in the supernatant (SN) was measured 48 h postinfection.C, Western blot detection of PARP cleavage in mutant and Ad5-infected A549 cells. Mock-infected cells or infected with Ad5, AdT1, or AdE3/19K-445A or treated with0.075 Amol/L and 0.2 Amol/L of staurosporine (SSP ) as a positive control of apoptosis induction were analyzed. The cell lysates were processed for detection ofPARP cleavage at 16, 24, 40, and 48 h postinfection. D, viral release and production of Ad5 and AdE3/19K-445A in carcinoma-associated fibroblasts in the presenceof a broad spectrum apoptosis inhibitor, Q-VD-Oph. Cells were infected and incubated in the presence or absence of 20 Amol/L Q-VD-Oph. The extracellular (SN)and intracellular (CE) viral content was measured 48 h postinfection.





injected with AdT1 showed a more diffuse staining compared withAd5, suggesting an improved spread of E3/19K-445A mutant virus(Fig. 6D).

Discussion

Improving the antitumor potency of current oncolytic adenovi-ruses represents one of the major challenges to obtain systemictherapeutic effect in clinical trials. Bioselection of randomlymutagenized pools of human Ad5 by repeated passaging under apredefined set of conditions is a classic virology strategy that hasbeen recently postulated as a powerful method to develop morepotent adenoviruses (18, 19). We hypothesized that the passage ofAd5 random mutants in an in vivo murine model of human cancerwould provide a selective pressure environment closer to humantumors in the clinical setting, where the presence of a three-dimensional tumor stroma, the host immune system, and theproliferating status of tumor cells differ from those encountered inin vitro cell cultures.

By using this approach, we have isolated a new mutant virus,AdT1, which displays an enhanced antitumor activity wheninjected systemically in mice carrying s.c. human tumors. In vitrocharacterization revealed that AdT1 displays a large-plaquephenotype, which correlates with an increased cytotoxicity andrelease of viral progeny after replication, and a total viral yield notmodified with respect to Ad5. Interestingly, the phenotype is not

restricted to the pancreatic tumors where the bioselection processwas developed, but the increased progeny release is also evident indifferent tumor cell types, including melanoma, lung, head andneck, and colorectal adenocarcinomas. Previously describedadenoviral mutants with large-plaque phenotype include mutantsin E1B/19K, in which E1A-induced apoptosis is not blocked (39),and premature cell death may lead to a reduced viral production.On the other hand, mutants that overexpress the adenovirus deathprotein (14, 35) or that express COOH truncating forms of thei-leader protein (18, 19) also display large-plaque phenotype,improving viral release without affecting the total yield. A doublestrategy, including the generation of Ad5/AdT1 recombinants andcomplete genome sequencing, allowed us to conclude that thealteration responsible for the AdT1-phenotype was the insertion ofone nucleotide in the E3/19K coding region (mutation E3/19K-445A), subsequently leading to the COOH truncation of the proteinand loss of its ER retention signal. The E3/19K glycoprotein retainsthe class I MHC heavy chain in the ER (40, 41) and prevents tapasinprocessing of class I MHC-bound peptides (33). However, to date, ithas never been postulated as capable of modifying viral cyclekinetics and its functions are assumed to be nonessential foradenovirus growth in vitro (42), which has led to its deletion inmany oncolytic adenoviruses (2).

Our studies show that the translocation of E3/19K-445A to theplasma membrane is the determinant for the enhanced releasephenotype of AdT1 and AdE3/19K-445A. Other mutants that also

Figure 5. Adenovirus mutant AdE3/19K-445A enhances membrane permeabilization and increases intracellular Ca2+ concentration. A, kinetics of cell permeability ofAd5-infected and AdE3/19K-445A–infected A549 cells. Cells were infected and, at the indicated time points, trypsinized and incubated with PI. The number ofpermeable cells (PI-positive cells) was determined. The intensity of PI for infected cells is represented in a logarithmic scale. B, kinetics of cytosolic free Ca2+ duringAd5 and AdE3/19K-445A infection. A549 cells infected with Ad5 or AdE3/19K-445A were incubated at the indicated time points with Fluo-3 AM and Fura-red AM. TheFluo-3/Fura-red ratio for each cell was analyzed, and cells were divided into a population with a high ratio (high intracellular [Ca2+]) or a low ratio (low intracellular[Ca2+]). Percentage of cells with high intracellular calcium concentration is plotted for each time point. C, cell permeability of Ad5-infected and AdE3/19K-445A–infectedCAF1 cells. D, viral release and production of Ad5, AdT1, and AdE3/19K-445A in the presence or absence of extracellular Ca2+ in carcinoma-associated fibroblasts.After infection, cells were incubated in normal or Ca2+-free medium, and extracellular and intracellular viral contents were measured 48 h postinfection.

Cancer Research




induce a change of localization of E3/19K to the cell membrane bydifferent genetic alterations, such as AdE3/19K-KS, also phenocopythe enhanced release of AdT1. This rules out the possibility that theprecise E3/19K-445A alteration provokes a change in the complexsplicing balance of the E3 region, leading to differential expressionof ADP and enhanced cell death, as found with other E3 mutants(36). In fact, analysis of E3-specific mRNAs does not suggest anyimbalance between the levels of the different transcripts after AdT1infection with respect to Ad5. Furthermore, protein levels of ADP,E1A, and major structural proteins are not modified in E3/19K-445A containing adenoviruses. Overall, these data suggest that theviral cycle of AdT1 is only affected in its latest events, i.e., theplasma membrane permeabilization and the mutant virion release.

How E3/19K-445A relocalization improves adenovirus release isnot clear at the present. This function is not related to the ability ofE3/19K to bind MHC-I, its canonical target, but the presence of themutation confers the protein the ability to initiate a new pathway.Apoptosis induction seems not to be the basis of the enhancedspreading of the E3/19K-445A mutants, as is the case for otherlarge-plaque mutant adenoviruses (39). Interestingly, our resultsprovide evidence that infection with E3/19K-445A expressingadenoviruses disrupts intracellular Ca2+ gradients maintained bythe plasma and ER membranes, leading to an increase in thecytoplasmic calcium levels. Furthermore, the requirement forextracellular Ca2+ availability for the enhanced release phenotypeof AdT1 suggests that such increases in intracellular calcium aredue, at least in part, to the influx of extracellular calcium. Increasesin intracellular calcium have been shown to be involved in animalvirus-induced cytophatic effects and cell killing (43). Protein 2Bfrom coxsackievirus, a protein with viroporin function, graduallyinduces the influx of extracellular Ca2+, and release of Ca2+ from ERstores to permeabilize the plasma membrane and ultimately causemembrane lesions that allow the release of virus progeny (38).Interestingly, wild-type E3/19K is capable of mobilizing the ER poolof Ca2+ through an indirect mechanism (29). Although a direct rolein Ca2+ mobilization by E3/19K-445A warrants experiments withthe protein being autonomously expressed, we hypothesize that themutant E3/19K acts functionally as a viroporin. Viroporins arenonessential, virally encoded, small membrane proteins that formselective channels in membranes, allowing the diffusion of ions andsmall molecules. Their deletion affects mainly the assembly andexit of virions (37). Their functions are related to both themodification of cell permeability to ions and membrane destabi-lization. The use of viroporins to promote progeny release iscommon to several families of viruses, such as picornavirus (38, 44),retrovirus (45), and orthomyxoviruses (46). In this context, ourresults suggest that, as a consequence of the selective pressureexerted in human xenografted tumors in vivo , adenovirus hasevolved to incorporate an alternative and more efficient mecha-nism of release, which it lacks in its native form (where it uses ADPas the main mechanism), but that other animal viruses havepreviously selected. Moreover, the fact that the mutation selectedin the current adenovirus bioselection process affects one of themain limitations found in oncolytic viruses suggests that pressureencountered in the environment of human tumors in vivo may beparticularly relevant. Curiously, the latter stages in the viral cycleseem to be more prone to optimization within the tumor milieu.Further insight into the mechanism of release of E3/19K-COOHterminal truncated mutants could provide additional clues tounderstand the poorly characterized mechanism of release ofnative adenovirus modulation.

Figure 6. In vivo efficacy. A, NP-9 s.c. tumor xenografts were treated i.v. withPBS, Ad5, or AdE3/19K-445A. Points, mean percentage of tumor growth(n = 10); bars, SE. #, significant (P = 0.03) compared with the group treatedwith PBS; *, significant (P = 0.008) compared with mice injected with Ad5.B, NP-9 tumor xenografts were treated intratumorally with PBS, Ad5, AdT1, orAdE3/19K-445A. #, significant (P = 0.003) compared with tumors injected withPBS; *, significant (P = 0.002) compared with tumors injected with PBS. C,Syrian hamsters with s.c. HP-1 tumors were treated intratumorally with PBS,Ad5, or AdT1. Columns, mean percentage of tumor growth (n = 10); bars, SE.#, significant (P = 0.01) compared with tumors injected with PBS; *, significant(P = 0.04) compared with tumors injected with PBS. D, adenovirus intratumoralreplication assessed by adenovirus detection in HP-1 pancreatic tumors growns.c. in immunocompetent hamsters. Frozen sections of tumors obtained at day29 after intratumoral injection of PBS (a ), or 2.5 � 1010 vp of Ad5 (b ) or AdT1mutant (c ) were treated with an anti-adenovirus antibody and counterstainedwith 4¶,6¶-diaminidino-2-phenylindole (�200). Sections taken from tumorstreated with AdT1 displayed a more diffuse staining for adenovirus presencecompared with Ad5-injected ones (green dots ). N, necrotic areas.





In terms of the immune response modulation, the E3/19K-445Aform differs from its wild-type counterpart in its ability to bind andretain MHC class I molecules, but it is expected to retain thetapasin inhibition ability that wild-type E3/19K possesses. Mutantswith truncated COOH terminal forms of the protein showed theability to delay by 10-fold the maturation of class I moleculesbecause they are still able to interact with TAP and interfere withtapasin function (33), and T1 mutants should behave similarly.More recently, it has been reported that E3/19K is additionallyinvolved in evasion of natural killer (NK) cell recognition throughits ability to interact with MHC-I chain–related proteins A and B(47). Although such function is impaired for E3/19K-445A protein,enhanced antitumoral activity of AdE3/19K-445A mutants obtainedin different in vivo models in mice, where NK function is present,suggest that the lack of such function can be efficiently by-passed,at least in the particular environment associated to tumors.Further experiments in more immunocompetent models arenecessary to establish the exact immunologic implicationsassociated with the presence of the 445A mutation.

Transcription of E3 proteins is mainly regulated by E1Aexpression through ATF and AP-1 boxes within the E3 promoter(48), implying that the enhanced release phenotype associated withthe expression of the mutant version of E3/19K is restricted to cellswhere E1A activation is permitted. We are currently generatingconditional replicating adenoviruses that include the E3/19K-445Aprotein in the backbone of tumor promoter–driven E1A-regulatedadenoviruses, such as ICOVIR-5 (11, 49), with the aim of increasingits ability to spread intratumorally. Interestingly, the increase inprogeny release of AdT1 and AdE3/19K-445A mutants is speciallynoted in cancer-associated fibroblasts, a perpetually activated

fibroblast population within desmoplastic lesions that are associ-ated with malignant tumors (50). The release of E3/19K-445Amutant progeny occurs 3 days earlier with respect to that of Ad5 inthe same cell population. Tumor stroma, composed of theextracellular matrix and mesenchymal cells, has been consideredas a barrier to the efficient intratumoral diffusion of viruses(11, 12). Although the ability of oncolytic viruses to replicate intumor-associated fibroblasts will depend on the selectivitymechanism in which the virus is based, the proliferative statusof such fibroblasts (51) should allow efficient replication of pRB-dependent adenoviruses and the inclusion of E3/19K-445Amutation in ICOVIR-5 would provide the virus the ability to by-pass fibroblast-mediated barriers.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

Received 3/26/2008; revised 7/25/2008; accepted 8/25/2008.Grant support: Mutua Madrilena Medical Research Foundation (Spain), Departa-

ment d’Universitats, Recerca i Societat de la Informacio from Generalitat de Catalunyagrant 2005 SGR 00727, and European Union 6th framework project research contract18700 (Theradpox; R. Alemany). R. Alemany belongs to the Network of CooperativeResearch on Cancer (C03-10), Instituto de Salud Carlos III of the Ministerio de Sanidady Consumo, Government of Spain.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Eduard Serra, Blanca Luena for her technical assistance in animalexperiments, Dr. J.W. Yewdell (Laboratory of Viral Diseases, NIAID, NIH) and Dr.W.S.M Wold (St. Louis University) for providing Tw1.3. and anti-ADP antibodies,respectively, and Lynda Coughlan for the extensive revision of this manuscript.

References

1. Liu TC, Galanis E, Kirn D. Clinical trial results withoncolytic virotherapy: a century of promise, a decade ofprogress. Nat Clin Pract Oncol 2007;4:101–17.

2. Alemany R. Cancer selective adenoviruses. MolAspects Med 2007;28:42–58.

3. Aghi M, Martuza RL. Oncolytic viral therapies - theclinical experience. Oncogene 2005;24:7802–16.

4. Chu RL, Post DE, Khuri FR, Van Meir EG. Use ofreplicating oncolytic adenoviruses in combinationtherapy for cancer. Clin Cancer Res 2004;10:5299–312.

5. Hamid O, Varterasian ML, Wadler S, et al. Phase II trialof intravenous CI-1042 in patients with metastaticcolorectal cancer. J Clin Oncol 2003;21:1498–504.

6. Small EJ, Carducci MA, Burke JM, et al. A phase I trialof intravenous CG7870, a replication-selective, prostate-specific antigen-targeted oncolytic adenovirus, for thetreatment of hormone-refractory, metastatic prostatecancer. Mol Ther 2006;14:107–17.

7. Alemany R, Suzuki K, Curiel DT. Blood clearance ratesof adenovirus type 5 in mice. J Gen Virol 2000;81:2605–9.

8. Bernt KM, Ni S, Gaggar A, et al. The effect ofsequestration by nontarget tissues on anti-tumorefficacy of systemically applied, conditionally replicatingadenovirus vectors. Mol Ther 2003;8:746–55.

9. Muruve DA. The innate immune response to adeno-virus vectors. Hum Gene Ther 2004;15:1157–66.

10. Wilson JM. Adenoviruses as gene-delivery vehicles.N Engl J Med 1996;334:1185–7.

11. Cascallo M, Alonso MM, Rojas JJ, et al. Systemictoxicity-efficacy profile of ICOVIR-5, a potent andselective oncolytic adenovirus based on the pRBpathway. Mol Ther 2007;15:1607–15.

12. Sauthoff H, Hu J, Maca C, et al. Intratumoral spreadof wild-type adenovirus is limited after local injection ofhuman xenograft tumors: virus persists and spreads

systemically at late time points. Hum Gene Ther 2003;14:425–33.

13. Wein LM, Wu JT, Kirn DH. Validation and analysis ofa mathematical model of a replication-competentoncolytic virus for cancer treatment: implications forvirus design and delivery. Cancer Res 2003;63:1317–24.

14. Doronin K, Toth K, Kuppuswamy M, et al. Tumor-specific, replication-competent adenovirus vectors over-expressing the adenovirus death protein. J Virol 2000;74:6147–55.

15. van Beusechem VW, van den Doel PB, Grill J, PinedoHM, Gerritsen WR. Conditionally replicative adenovirusexpressing p53 exhibits enhanced oncolytic potency.Cancer Res 2002;62:6165–71.

16. Wildner O, Blaese RM, Morris JC. Therapy of coloncancer with oncolytic adenovirus is enhanced by theaddition of herpes simplex virus-thymidine kinase.Cancer Res 1999;59:410–3.

17. Bristol JA, Zhu M, Ji H, et al. In vitro and in vivoactivities of an oncolytic adenoviral vector designed toexpress GM-CSF. Mol Ther 2003;7:755–64.

18. Yan W, Kitzes G, Dormishian F, et al. Developingnovel oncolytic adenoviruses through bioselection.J Virol 2003;77:2640–50.

19. Subramanian T, Vijayalingam S, Chinnadurai G.Genetic identification of adenovirus type 5 genes thatinfluence viral spread. J Virol 2006;80:2000–12.

20. Villanueva A, Garcia C, Paules AB, et al. Disruption ofthe antiproliferative TGF-h signaling pathways inhuman pancreatic cancer cells. Oncogene 1998;17:1969–78.

21. Batra SK, Metzgar RS, Worlock AJ, Hollingsworth MA.Expression of the human MUC1 mucin cDNA in ahamster pancreatic tumor cell line HP-1. Int J Pancrea-tol 1992;12:271–83.

22. Alemany R, Curiel DT. CAR-binding ablation doesnot change biodistribution and toxicity of adenoviralvectors. Gene Ther 2001;8:1347–53.

23. Williams JF, Gharpure M, Ustacelebi S, McDonald S.Isolation of temperature-sensitive mutants of adenovi-rus type 5. J Gen Virol 1971;11:95–101.

24. Cascallo M, Gros A, Bayo N, et al. Deletion of VAI andVAII RNA genes in the design of oncolytic adenoviruses.Hum Gene Ther 2006;17:929–40.

25. Cox JH, Bennink JR, Yewdell JW. Retention ofadenovirus E19 glycoprotein in the endoplasmic retic-ulum is essential to its ability to block antigenpresentation. J Exp Med 1991;174:1629–37.

26. Tollefson AE, Scaria A, Saha SK, Wold WS. The11,600-MW protein encoded by region E3 of adenovirusis expressed early but is greatly amplified at late stagesof infection. J Virol 1992;66:3633–42.

27. Burchiel SW, Edwards BS, Kuckuck FW, et al. Analysisof free intracellular calcium by flow cytometry: multi-parameter and pharmacologic applications. Methods2000;21:221–30.

28. Jackson MR, Nilsson T, Peterson PA. Retrieval oftransmembrane proteins to the endoplasmic reticulum.J Cell Biol 1993;121:317–33.

29. Pahl HL, Sester M, Burgert HG, Baeuerle PA.Activation of transcription factor NF-nB by the adeno-virus E3/19K protein requires its ER retention. J Cell Biol1996;132:511–22.

30. Tollefson AE, Ryerse JS, Scaria A, Hermiston TW,WoldWS. The E3–11.6-kDa adenovirus death protein (ADP) isrequired for efficient cell death: characterization of cellsinfected with adp mutants. Virology 1996;220:152–62.

31. Hermiston TW, Tripp RA, Sparer T, Gooding LR,Wold WS. Deletion mutation analysis of the adenovirustype 2 E3–19K protein: identification of sequenceswithin the endoplasmic reticulum lumenal domain thatare required for class I antigen binding and protectionfrom adenovirus-specific cytotoxic T lymphocytes.J Virol 1993;67:5289–98.

32. Sester M, Burgert HG. Conserved cysteine residueswithin the E3/19K protein of adenovirus type 2 are

Cancer Research






essential for binding to major histocompatibilitycomplex antigens. J Virol 1994;68:5423–32.

33. Bennett EM, Bennink JR, Yewdell JW, Brodsky FM.Cutting edge: adenovirus E19 has two mechanisms foraffecting class I MHC expression. J Immunol 1999;162:5049–52.

34. White E, Grodzicker T, Stillman BW. Mutations inthe gene encoding the adenovirus early region 1B19,000-molecular-weight tumor antigen cause thedegradation of chromosomal DNA. J Virol 1984;52:410–9.

35. Tollefson AE, Scaria A, Ying B, Wold WS.Mutations within the ADP (E3-11.6K) protein alterprocessing and localization of ADP and the kineticsof cell lysis of adenovirus-infected cells. J Virol 2003;77:7764–78.

36. Scaria A, Wold WS. Fine-mapping of sequences thatsuppress splicing in the E3 complex transcription unit ofadenovirus. Virology 1994;205:406–16.

37. Gonzalez ME, Carrasco L. Viroporins. FEBS Lett2003;552:28–34.

38. van Kuppeveld FJ, Hoenderop JG, Smeets RL, et al.Coxsackievirus protein 2B modifies endoplasmicreticulum membrane and plasma membrane perme-

ability and facilitates virus release. EMBO J 1997;16:3519–32.

39. Pilder S, Logan J, Shenk T. Deletion of the geneencoding the adenovirus 5 early region 1b 21,000-molecular-weight polypeptide leads to degradation ofviral and host cell DNA. J Virol 1984;52:664–71.

40. Andersson M, Paabo S, Nilsson T, Peterson PA.Impaired intracellular transport of class I MHC antigensas a possible means for adenoviruses to evade immunesurveillance. Cell 1985;43:215–22.

41. Burgert HG, Kvist S. An adenovirus type 2 glycopro-tein blocks cell surface expression of human histocom-patibility class I antigens. Cell 1985;41:987–97.

42. BhatBM,WoldWS. A small deletiondistant froma spliceor polyadenylation site dramatically alters pre-mRNA pro-cessing in region E3 of adenovirus. J Virol 1987;61:3938–45.

43. Chami M, Oules B, Paterlini-Brechot P. Cytobiolog-ical consequences of calcium-signaling alterationsinduced by human viral proteins. Biochim Biophys Acta2006;1763:1344–62.

44. de Jong AS, Schrama IW, Willems PH, et al. Multi-merization reactions of coxsackievirus proteins 2B, 2Cand 2BC: a mammalian two-hybrid analysis. J Gen Virol2002;83:783–93.

45. Gonzalez ME, Carrasco L. The human immunodefi-ciency virus type 1 Vpu protein enhances membranepermeability. Biochemistry 1998;37:13710–9.

46. Guinea R, Carrasco L. Influenza virus M2 proteinmodifies membrane permeability in E. coli cells. FEBSLett 1994;343:242–6.

47. McSharry BP, Burgert HG, Owen DP, et al. AdenovirusE3/19K promotes evasion of NK cell recognition byintracellular sequestration of the NKG2D ligands majorhistocompatibility complex class I chain-related pro-teins A and B. J Virol 2008;82:4585–94.

48. Akusjarvi G. Proteins with transcription regulatoryproperties encoded by human adenoviruses. TrendsMicrobiol 1993;1:163–70.

49. Alonso MM, Cascallo M, Gomez-Manzano C, et al.ICOVIR-5 shows E2F1 addiction and potent antigliomaeffect in vivo . Cancer Res 2007;67:8255–63.

50. Kalluri R, Zeisberg M. Fibroblasts in cancer. Nat RevCancer 2006;6:392–401.

51. Hasebe T, Sasaki S, Imoto S, Ochiai A. Proliferativeactivity of intratumoral fibroblasts is closely correlatedwith lymph node and distant organ metastases ofinvasive ductal carcinoma of the breast. Am J Pathol2000;156:1701–10.



2008;68:8928-8937. Cancer Res Alena Gros, Jordi Martínez-Quintanilla, Cristina Puig, et al. Adenovirus 5 Release and Improves Its Antitumoral PotencyBioselection of a Gain of Function Mutation that Enhances

Updated version

http://cancerres.aacrjournals.org/content/68/21/8928

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2008/10/27/68.21.8928.DC1.html

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/68/21/8928.full.html#ref-list-1

This article cites 51 articles, 22 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/68/21/8928.full.html#related-urls

This article has been cited by 2 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

[email protected] at

To request permission to re-use all or part of this article, contact the AACR Publications


http://cancerres.aacrjournals.org/content/68/21/8928

http://cancerres.aacrjournals.org/content/suppl/2008/10/27/68.21.8928.DC1.html

http://cancerres.aacrjournals.org/content/68/21/8928.full.html#ref-list-1

http://cancerres.aacrjournals.org/content/68/21/8928.full.html#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

mailto:[email protected]


Documents

Bioselection of a Gain of Function Mutation that Enhances Adenovirus 5 Release and Improves Its Antitumoral Potency